Skip to main content

Simulation run on this device. Scenario values are not sent anywhere — the engine is fully local.

research studio · pro

Paper-to-Code Repro Checker

Score a paper's reproducibility surface: code, seeds, splits, configs.

Is this result reproducible from what's published?

Engine 1.0.0 · studio
Reproducibility
0/10
Signals found
0/5
Paper length
0 words
Code / weights released
absent
Hyperparameters
absent
Dataset & splits
absent
Seeds & environment
absent
Eval protocol
absent
  • medium
    Missing: Code / weights released

    no code/weights availability signal

  • medium
    Missing: Hyperparameters

    learning rate / batch size / optimizer not specified

  • medium
    Missing: Dataset & splits

    dataset identity or train/val/test split unclear

  • medium
    Missing: Seeds & environment

    seeds / hardware / software versions absent

  • medium
    Missing: Eval protocol

    metric definition or baselines missing

Each missing dimension is a guess a reproducer would have to make; the first one is usually the highest-value fix.

Method

  • Scores five reproducibility dimensions 0–2 from explicit signals in the text: code/weights, hyperparameters, dataset & splits, seeds/environment, and eval protocol.
  • A missing dimension is a guess a reproducer would have to make — the score is a checklist, not a promise.